Edinburgh Explorer Switching the orientation of Jahn-Teller axes in oxime-based Mn(III) dimers and its effect upon magnetic exchange

A family of Mn III dimers of general formula [Mn III 2 (R-sao) 2 (dpa) 2 ](ClO 4 ) 2 ( 1 - 5 ) has been synthesised using derivatised phenolic oximes (R-saoH 2 , where R = H, Me, Et, Ph) in combination with di-(2-picolyl)-amine (dpa). Their structures reveal a double-oxime bridged [Mn III (NO)] 2 magnetic core in which the Jahn-Teller axes lie perpendicular to the bridging plane, in contrast to two previously reported family members ( 6 , 7 ). The switch in the orientation of the Jahn-Teller axes is enforced through the use of the chelating ligand which is present in 1 - 5 and absent in 6 - 7 . Dc magnetic susceptibility measurements reveal that the exchange interactions between the Mn III metal centres in 1 - 5 are antiferromagnetic in contrast to that observed for 6 and 7 which are ferromagnetic. DFT calculations performed on complexes 1 - 6 reproduce both the sign and strength of the J values found experimentally. Molecular orbital analysis unlocks a common mechanism of magnetic coupling based upon the orientation of the Jahn-Teller axis, with the magneto-structural correlation also dependent upon the Mn-N-O-Mn angles – with ferromagnetic interactions at smaller dihedral angles.


Introduction
Molecular Magnets -molecules containing exchange-coupled paramagnetic metal ions -are an enormously important class of material with potential application across a diverse range of fields from information storage, quantum computation and molecular spintronics to cryogenic refrigeration and biomedical imaging. [1][2][3][4][5] Such species combine the macroscopic properties of magnets with the quantum mechanical properties of molecules, allowing detailed study of quantum size effects in monodisperse, reproducible, orientable, chemically tuneable molecules. 6 Recent years have witnessed a plethora of important scientific breakthroughs including the observation of quantum tunnelling of the magnetization, quantum phase interference, memory effects at T = 14 K, single-molecule magnetic detection, the construction of molecular spintronic devices, molecules and coordination polymers displaying enormously enhanced magnetocaloric effects, and the development of protocols for molecule-based quantum information processing. [7][8][9][10][11][12][13][14][15][16][17][18][19][20] The discovery of such phenomena, and any potential exploitation thereof, is based upon a detailed understanding of, and subsequent control over, the structure-magnetism relationship. This in turn is derived from a systematic synthetic study of the effects of ligand design, metal identity and reaction conditions upon symmetry and structure, in tandem with detailed physical characterisation and theoretical analysis. 21 Manganese is often a good choice for constructing such molecules: it commonly exists in the II+, III+ and IV+ oxidation states, resulting in the majority of clusters being mixed-valent, which, in turn, decreases the likelihood of antiferromagnetically coupled cages possessing diamagnetic spin ground states. 6 The Jahn-Teller distorted Mn III ion also provides a source magnetic anisotropy that has often been employed in the construction of Single-Molecule Magnets (SMMs), 6 whilst the isotropic Mn II ion finds application in clusters displaying an enhanced magnetocaloric effect (MCE), a phenomenon that can be exploited for low temperature cooling. 17 Application first requires the development of magneto-structural correlations, but this can often be a difficult task due to prohibitively large nuclearities and/or complex exchange interaction patterns. One way to overcome such issues is to construct families of low nuclearity cages (e.g. dimers, trimers, tetramers) which mimic the building blocks of the bigger cages, and investigate how small structural perturbations affect magnetic exchange. Detailed studies on Mn III dimers, however, remain relatively scarce. 22 A recent experimental and theoretical study of alkoxide-bridged [Mn III (OR)]2 dimers revealed that the dominant structural feature controlling magnetic exchange (J) was the relative orientations of the Jahn-Teller axes. 22 A parallel orientation, perpendicular to the bridging plane of the molecule (Type I, Figure 1) resulted in weak antiferromagnetic exchange; a parallel orientation within the bridging plane (Type II, Figure  1) led to borderline cases in which the exchange could be either weakly ferromagnetic or weakly antiferromagnetic; whilst a perpendicular orientation (Type III, Figure 1) produced ferromagnetic exchange. For the oxime/oxo bridged species [Mn III 2(O)(NO)] (Figure 1, A) the Mn-N-O-Mn torsion angle dominates, with a linear correlation found between the magnitude of J and the torsion angle. 22 DFT studies on doubleoxime bridged dimers [Mn III (NO)]2 ( Figure 1, B) also revealed a pronounced dependence of the exchange coupling on the relative twisting of the oxime moiety, as proposed previously in more complicated [Mn III 3] and [Mn III 6] clusters, 23 resulting from an accidental orthogonality between the Mn-N-O plane of the first Mn III ion and the Jahn-Teller axis of the second Mn III ion. 22 In order to extend the family of double-oxime bridged Mn III dimers -which has just two members -we herein report a series of double-oxime bridged [Mn III (NO)]2 species in which employment of the chelating ligand di-(2-picolyl)-amine, dpa ( Figure 2), switches the orientation of the Jahn-Teller axes from being in the bridging plane (structure type B, Figure 1) to being perpendicular to the bridging plane (structure type C, Figure 1) and present a combined experimental and theoretical analyses of the effect of this Jahn-Teller switching upon the magnetic exchange between the two Mn III ions.

Materials and Physical Measurements
All synthetic procedures were performed under aerobic conditions using chemicals as received (reagent grade). Caution! Although no problems were encountered here, care should be taken when handling the potentially explosive perchlorate anion. The substituted phenolic oximes were synthesised using the appropriate precursor ketones, hydroxylamine hydrochloride and sodium acetate in EtOH, as described in the literature. 24 Variable temperature, solid state magnetic susceptibility data down to 5 K were collected on a Quantum Design MPMS XL SQUID magnetometer equipped with a 7 T dc magnet. Diamagnetic corrections were applied to the observed paramagnetic susceptibilities using Pascal's constants.

Synthesis
General procedure for the synthesis of [Mn III 2(R-sao)2(dpa)2](ClO4)2: Mn(ClO4)2.6H2O (0.4 mmol) and the appropriate phenolic oxime (0.4 mmol) were dissolved in MeOH (10 mL). NH4OH (1.6 mmol) was then added to the solution and the reaction left to stir for 40 minutes at room temperature. dpa (0.4 mmol) in EtOH (10 mL) was added to the reaction mixture, which was then filtered after 5 minutes of additional stirring. The black mother liquor was left to evaporate over a period of 4 days, during which time black block-like crystals grew in varying yields. See the SI for full synthetic details.

X-Ray Crystallography
Diffraction data for samples 1 and 3-5 (CCDC numbers 1044300, 1044302-1044304) were collected on an Oxford Diffraction SuperNova diffractometer using Mo or Cu Kα radiation. The crystal temperature was maintained at 120 K Oxford Cryosystems Cryostream 700+ low temperature device. The structures were solved by direct methods and refined by full-matrix least-squares techniques on F 2 using the programs SHELXL 25a and Olex2. 25b X-ray data of complex 2 (CCDC 1044301) were collected at 100 K (Oxford Cryosystems Cobra) on a Rigaku AFC12 goniometer equipped with an enhanced sensitivity (HG) Saturn724+ detector mounted at the window of an FR-E+ SuperBright molybdenum rotating anode generator with VHF Varimax optics (70m focus). The crystal structure was solved by charge flipping methods in SUPERFLIP 25c and the fullmatrix least-squares refinement on Fo 2 was carried out using SHELXL. 25a All non-hydrogen atoms were refined with anisotropic displacement parameters. All hydrogen atoms were added at calculated positions and refined using a riding model with isotropic displacement parameters based on the equivalent isotropic displacement parameter (Ueq) of the parent atom. Table S1 in the supplementary information contains a summary of the experimental crystallographic parameters.

Computational details
Calculations were performed on the full crystal structures of 1-6. A hybrid B3LYP 26 functional with TZV basis set 27 was employed for calculating the exchange constants as implemented in the G09 28 suite of programs. The high spin states (E-HS) and low spin states (E-BS) were estimated using single determinant wave functions and the broken symmetry approach, 29 respectively, and the corresponding J values were computed from the difference between E-HS and E-BS. Further details about the computational methodology are discussed elsewhere. 30

SQUID Magnetometry
Dc magnetic susceptibility measurements were performed on polycrystalline samples of 1-3 and 5 in the temperature range T = 5 -300 K in an applied magnetic field of B = 0.1 T. The poor yield of complex 4 precluded measurement. The experimental results are shown in Figure 4 in the form of χMT products vs T, where χM is the molar magnetic susceptibility, along with the data for complexes 6 and 7. 22 At the highest temperatures measured the χMT values for 1-3 and 5 are ~4.1, 4.7, 4.8 and 5.0 cm 3 K mol -1 , respectively, all somewhat lower than that expected for two non-interacting, high spin d 4 ions with g = 2 (6 cm 3 K mol -1 ). Upon cooling the χMT products of each sample behave very similarly, slowly decreasing with decreasing temperature reaching values of ~0.2 (1), 0.7 (2), 0.2 (3) and 1.1 cm 3 K mol -1 (5) at T = 5 K, indicative of the presence of weak antiferromagnetic exchange between the two Mn III ions. This is in stark contrast to that observed for complexes 6 and 7 which display weak ferromagnetic exchange, as discussed in the articles highlighted in reference 22.
The data can be fit ( Figure 4, Table 2) to the model described by isotropic spin-Hamiltonian (1), where the indices i and j refer to the two Mn III centres, J is the isotropic exchange interaction parameter, Ŝ is a spin operator, μB is the Bohr magneton, B is the applied magnetic field vector and g =1.98 is the g-factor of the Mn III ions. The best-fit parameters are J = -5.73 cm -1 (1), -3.63 cm -1 (2), -5.63 cm -1 (3) and -2.05 cm -1 (5). The ground spin-state of all four complexes is a spin singlet (S = 0; energy versus spin state plots are given in the SI, Figures S6-9). This can be compared to the S = 4 ground states observed for complexes 6 and 7 resulting from exchange interactions of +2.20 and +1.24 cm -1 , respectively. The difference between the magnetic behaviour of 1-3, 5 and 6-7 is intriguing, particularly given they have the same [Mn(NO)]2 magnetic core, and thus it would appear that the orientation of the Jahn-Teller axes is an important structural parameter. In order to examine this in more detail we have therefore turned to Density Functional Theory.   Theoretical studies have been carried out on complexes 1-6 in order to evaluate the exchange parameters and to reveal the mechanism of exchange. The DFT computed J values for complexes 1-6 are -5.76 cm -1 , -2.16 cm -1 , -5.15 cm -1 , -3.66 cm -1 , -2.99 cm -1 and +2.65 cm -1 , respectively (Table 1). Complexes 1-5 are found to exhibit weak antiferromagnetic exchange interactions, while complex 6 exhibits ferromagnetic exchange. Calculations reproduce not only the sign of J accurately but also their magnitude and the trend in magnitudes compared to that observed experimentally. To understand the electronic reasons behind the variation in the nature and magnitude of exchange and to analyse the effect of the relative orientation of the Jahn-Teller axis (structure types B and C), molecular orbitals (MO) and overlap integrals have been analysed for complexes 1-6. The net exchange interaction in the dinuclear Mn III moiety has two contributions: (i) an antiferromagnetic JAF contribution arising solely from overlap between the singly occupied MOs (SOMOs) of the Mn III ions, and (ii) a ferromagnetic JF contribution arising from the orthogonality of the SOMOs (negligible overlap) and from the crossinteraction 22b,c between the occupied and the empty d-orbitals (here between the dx 2 -y 2 and all the other Mn III d-orbitals). Amongst the structures studied, the major contributor to the JAF term is the overlap between the dxy orbitals via the oxime bridge. This contribution is significant for complexes 1-5 possessing type C structures, as indicated in Figure 1 and Figure 5a. The second dominant contribution to the JAF term is the dxz-dxz overlap, also routed through the Mn-N-O-Mn bridge. Thus for complexes 1-5 the Mn-N-O-Mn torsion angle clearly dictates the strength of orbital overlap and thus the J value. These two overlaps are expected to be minimal in complex 6, resulting in a very small JAF term. However as the Jahn-Teller axes are now oriented along the oxime bridges, this allows for efficient dz 2 -dz 2 overlap (see Figure 5b). In this scenario this latter overlap is dominant and this contributes to the JAF term. As shown in Figure 5b, this overlap is also strongly dependent upon the Mn-N-O-Mn torsion angle, with negligible overlap expected when the torsion angle approaches zero. Apart from these three prominent interactions, other orbital overlap values are small suggesting orthogonality between the SOMOs and a contribution to the JF term. Along with moderate cross- interactions (dx 2 -y 2 -dxz) this leads to a significant JF term for all complexes 1-6. The sign of J in each case is dictated by the dominant factor; for 1-5 the two strong overlaps overwhelmingly dominate leading to net antiferromagnetic exchange, while the prominent dz 2 -dz 2 overlap observed in complex 6 is overshadowed by the JF contributions leading to a net ferromagnetic interaction.
Spin density plots were computed in order to understand the origin of the electronic differences in complexes 1-6. Unpaired electrons in t2g orbitals usually favour a spin polarisation mechanism, whereas those in the eg orbitals tend to facilitate a spin delocalisation mechanism. In the case of the Mn III ion a mixture of these two mechanisms is found to operate since the ion possesses three unpaired electrons in the t2g orbitals and one unpaired electron in the eg orbital. The spin density of the Mn III ions in 1-6 is found to be <3.8 which shows that it is centred on the metal ion with a dominant spin delocalisation mechanism. The bridging N and O spin densities are different for the two structure types studied; for complex 1 all the bridging N and O atoms possess negative spin density indicating spin polarisation, while complex 6 possesses positive spin density on the O atom due to spin delocalistion as it lies along the Jahn-Teller axis (see Table S8 in ESI and Figure 6) and negative spin density on the N atom due to spin polarisation.  Figure 7. Our correlation reveals that J switches its sign from antiferromagnetic to ferromagnetic with decreasing dihedral angle. As the torsion angle decreases, the cross-interaction between dx 2 -y 2 and dxz/dxy strengthens, leading to a larger JF contribution which in turn results in ferromagnetic exchange. Our calculations predict that a more planar Mn-N-O-Mn moiety of structure type C will exhibit ferromagnetic exchange coupling. Synthetic efforts to produce such a molecule are underway in our laboratory.

Conclusions
The simple reaction between a Mn salt, a phenolic oxime and the tridentate chelate dpa affords a small family of double-oxime bridged Mn III dimers, whose structures differ from previous examples in the orientation of their Jahn-Teller axes. Previous family members conform to structure type B with the JT axes coparallel and in the bridging plane between the Mn III ions, resulting in ferromagnetic exchange. The addition of a chelating ligand results in the novel complexes 1-3, 5 whose JT axes have now switched orientation giving structure type C, where they lie perpendicular to the [Mn(ON)]2 bridging plane. The result is that the magnetic exchange between the metal centres becomes weakly antiferromagnetic. The employed DFT methodology is able to reproduce both the sign and magnitude of the exchange interaction in complexes 1-6. MO analysis reveals two dominant overlaps between the dxy-dxy and dxz-dxz orbitals which are found to control the sign and magnitude of exchange in complexes 1-5.
In the case of complex 6, strong dz 2 -dz 2 overlap is overcome by cross-interactions and orbital orthogonality resulting in a weak ferromagnetic interaction. The general mechanism of exchange proposed can, in principle, be extended to any double oxime bridged Mn III dimer. Our developed magneto-structural correlation suggests a switch in the sign of the exchange interaction from antiferromagnetic to ferromagnetic upon decreasing the Mn-N-O-Mn torsion angle to approximately 20º.